Thursday’s column elicited quick a bit of feedback – most of it, it has to said, suggesting that I had wasted my time writing it! The strong consensus out there is that the zero brokerage model will not work in foreign exchange. In my defence, I was actually asking the question rather than answering it and I did also note in the column the argument put forward by many of you – that the different market structures are a big stumbling block.
I am therefore, back in my box…and if you believe that you’ll believe anything!
I should reiterate before I go any further that I remain sceptical any platform would make such a move, however that is not the same as suggesting it will not work. Developing the theme further in my mind it really seemed clear to me that if it was one or both of the primary venues that kicked off such a price war then the impact could be quick and deep as far as the industry is concerned. An argument is often put forward that the primary venues are the least efficient thanks to the baggage and legacy technology they have built up over the years – and that is valid. I would also point out, however, that there are plenty of venues with lower volumes that are also reliant upon their brokerage income – if that disappears, they have nothing to fall back upon.
Several of you also asked the very valid question, ‘how would a firm make money in a zero brokerage environment?’ I think I covered it a little in Thursday’s column but basically it would be via ancillary services – firm and foremost data, and this is where the primary venues have the advantage. As I noted, last year a paper by the BIS Markets Committee observed that while volumes on the primary venues were going down, their importance to price formation was as important as ever, perhaps more so. This means that if these firms cut their brokerage to zero, as the US equity brokers have, they would still have revenue from their very important data businesses. Equally, zero brokerage would, one would think in this day when every cent is analysed to the nth degree, see increased volume on these platforms, leading to more data. More importantly it would quickly damage those platforms that have been eroding their business over the past five years.
Another factor that makes me think this could be the time for such a move is, and I wrote about this recently, the feeling in the FX world that there are too many platforms out there. Not only would market participants welcome the serious revenue boost from no brokerage on trading, they could also trim a few dollars off the budget by cutting the odd connection.
For those venues that would be threatened by such a move it is interesting to ponder their options in such an environment. The data route is a possible, but given how sources are now telling me that the data they get from so many venues is merely a reiteration of the primary venues, is that valuable enough for a trading business to maintain a connection? Probably not, and the fact that many of these venues rely heavily upon liquidity recyclers, means their data will be relatively worthless. The evangelists will have us believe otherwise, but you can have too much data.
Another opportunity could be to charge one side of the trade, but is this fair? I have a problem with platforms that currently charge one party only (normally the LP) because it is reinforcing the fallacy that liquidity is plentiful and easy to supply, so would not like to see this model have a broader adoption. A side issue to this, one that the US equity brokerages no doubt hope to exploit, is the venues paying the brokerages for market order flow, but this model is under regulatory pressure already and is probably not long for this world. Such a model could keep the LPs onside and create the data to be sold, that’s true; but it will be shut, and locked, pretty soon.
One avenue for the threatened venues to explore is to fall back upon their wealthy owners, if they have them. The FX business can be merged into the other asset classes supported by the owner (which, let’s face it, is usually an exchange) which means costs can be cut (at the cost, probably, of a dilution in service quality). Again though, I have my doubts, mainly because so few organisations in any segment have really been able to build a unified and profitable listed and OTC business unit.
Finally, if these venues have good technology, there may be an opportunity to white label it to regional players – again though, this is likely to be a limited market and one that is saturated pretty quickly. The only answer would appear to be cut costs drastically to the stage at which you can at least turn a profit – it’s really just a question of whether the service quality is sufficient to keep the all-important customers on board.
Having noted all of this, of course, it could also be argued that such a dramatic shift in the landscape would have no effect because many of these other venues rely upon different liquidity providers to the top tier (the recyclers predominently), and these LPs could, theoretically, take the market data from the primary venue, trade there when they want to be a liquidity consumer (for free) and still be an LP on the other venue – such is the complexity of the modern FX market structure. This is a sound argument, but I think the fact that liquidity consumers would know that one of the key factors in this service would be last look, and that they would be paying to trade on non-firm liquidity, as opposed to not paying to trade on firm, would see them also gyrate in the direction of the primary venues.
It is important to note that I don’t think the FX market should revert back to the 1990s when there were the two primary venues and not a lot else, rather my argument is that the anonymous ECN model could maybe do with a few less participants. Look at the FX turnover surveys and listen to market participants, especially LPs, and you will see that the anonymous channel is losing popularity.
The Global FX Committee would also probably breathe a sigh of relief if the anonymous model was trimmed down in terms of venues, because if there is one area that seems to be a sticking point on industry reform it is how to enforce such change in an anonymous environment? We all know how we would like people to behave, but how do you check they are in an anonymous, fragmented, environment? Ask a platform provider – and I think this segment is under pressure to do more about the issue – and they will tell you there is not a lot they can do. Suppose then, that this attitude, allied to the demand for a simplification of what former Bank of England chief dealer Martin Mallett used to call the “spaghetti of liquidity”, persists in an environment in which one or more primary venues are not charging brokerage? It could be yet another strike against them.
So yes, I totally accept the wholesale scepticism with which my scenario was received, but no, I do not think this is a simple case of ‘it couldn’t happen here’. It would take a brave decision on the part of a primary venue, but perhaps the time for such a decision is approaching? The volume numbers are not good, and they do not seem to be improving. Since January 2014, so 57 months, there have been just 10 occasions when the ADV of EBS and Reuters/Refinitiv was above the $100 billion mark at both platforms (and four of those were in 2014). Prior to that, going back as far as records are available (2007 for EBS, 2009 for Reuters) there were only 10 months in which both platforms did not hit the same mark.
The business model probably needs to change and, assuming the LSE deal for Refinitiv goes through, both venues will have larger owners who may be willing to take a risk on a disruptive strategy. The LSE deal is very much about the data anyway and CME Group is a very profitable business, so if not now (or in the next year or so being more realistic) when?